联邦多任务学习

May, 2017

Federated Multi-Task Learning

Virginia Smith, Chao-Kai Chiang, Maziar Sanjabi, Ameet Talwalkar

TL;DR本文研究了在分布式设备网络上训练机器学习模型所面临的统计和系统挑战，提出了一种适用于多任务学习的优化方法MOCHA，并在考虑通信成本、慢设备和容错性等实际系统问题的前提下，实现了比替代方法更显著的加速效果。

Abstract

federated learning poses new statistical and systems challenges in training machine learning models over distributed networks of devices. In this work, we show that multi-task learning is naturally suited to hand